Prevalence of a disease plays an important role in your probability of having COVID-19 given you tested positive

The prevalence of a disease plays an important role in your probability of having it given you test positive.
statistics
uncertainty
coronavirus
casual inference
bayes theorem
Author

Lucy D’Agostino McGowan

Published

April 13, 2020

The prevalence of a disease plays an important role in your probability of having it given you test positive.

This is less relevant for testing positive for SARS-CoV-2 infection since RT-PCR rarely (if ever) has false positives. However, this may be relevant for antibody tests, which are less precise. Here is a quick explainer for how this works.

The numbers in this explainer are made up. Are some real numbers for the sensitivity and specificity (as it stands today): r tufte::margin_note("1. https://www.mayoclinicproceedings.org/article/S0025-6196(20)30365-7/fulltext <br> 2. https://www.fda.gov/media/136151/download <br> 3. https://www.centerforhealthsecurity.org/resources/COVID-19/serology/Serology-based-tests-for-COVID-19.html")

RT-PCR:
r emo::ji("point_right") sensitivity: can be as low as 70%, maybe closer to 90%¹
r emo::ji("point_right") specificity: ~100%²

Antibody test (depends on the test, the one approved in US³):
r emo::ji("point_right") sensitivity: 93.8%
r emo::ji("point_right") specificity: 95.6%